TUM Logo

Han Xiao

Dr. Han Xiao

Alumni

E-Mail:
now at Zalando, Data Scientist

My research aims at employing geometry and optimization methodologies to analyze the vulnerability of learning algorithms in adversarial settings. The goal is to develop robust learning algorithms that are resilient to adversarial noise.

Also, I maintain an interest in graphical modeling approaches to text mining, information retrieval and computer vision.

>> go to my personal website for more info

Publications

2014 Support Vector Machines under Adversarial Label Contamination
2013 Online Lazy Gaussian Process Committee and its Application in Real-Time Trajectory Prediction
Efficient Online Sequence Prediction with Side Information
Lazy Gaussian Process Committee for Real-Time Online Regression
Learning from Multiple Observers with Unknown Expertise
OPARS: Objective Photo Aesthetics Ranking System
2012 Adversarial Label Flips Attack on Support Vector Machines
Evasion Attack of Multi-Class Linear Classifiers
2011 A Supervised Topic Transition Model for Detecting Malicious System Call Sequences
2010 Toward Artificial Synesthesia: Linking Images and Sounds via Words
Efficient Collapsed Gibbs Sampling For Latent Dirichlet Allocation